[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7457.1306203521@localhost>
Date: Mon, 23 May 2011 22:18:41 -0400
From: Valdis.Kletnieks@...edu
To: Youquan Song <youquan.song@...el.com>
Cc: linux-kernel@...r.kernel.org, david.woodhouse@...el.com,
akpm@...ux-foundation.org, mingo@...e.hu, tglx@...utronix.de,
hpa@...or.com, hpa@...ux.intel.com, allen.m.kay@...el.com,
suresh.b.siddha@...el.com, rajesh.sankaran@...el.com,
asit.k.mallick@...el.com, kent.liu@...el.com,
Youquan Song <youquan.song@...ux.intel.com>
Subject: Re: [PATCH v3] x86, vt-d: enable x2apic opt out
On Mon, 23 May 2011 22:32:28 +0800, Youquan Song said:
> + no_x2apic_optout [Default Off]
> + With this option BIOS x2APIC opt-out request will be
> + ignored.
> + else if (!x2apic_supported() && cpu_has_x2apic)
> + WARN(1, "Your BIOS is broken and requested that x2apic be "
> + "disabled.\n This will leave your machine vulnerable to"
> + " irq-injection attacks\n"
> + "Use 'intel_iommu=no_x2apic_optout' to override BIOS "
> + "request\n");
If we're doing a WARN level here, what are the downsides of just automagically
forcing it rather than making them use a kernel parameter and reboot? Will
some systems fail to boot because the BIOS was in fact right in requesting
hat x2apic be turned off?
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists